JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1993, p. 3231-32390095-1137/93/123231-09$02.00/0Copyright X) 1993, American Society for Microbiology

Vol. 31, No. 12

Clinical Significance, Biochemical Features, and SusceptibilityPatterns of Sporadic Isolates of the Mycobacterium

chelonae-Like OrganismRICHARD J. WALLACE, JR.,l* VELLA A. SILCOX,2 MICHIO TSUKAMURA,3t

BARBARA A. BROWN,' JAMES 0. KILBURN,2 W. RAY BUTLER,2 AND GRACE ONYI'Department of Microbiology, The University of Te-xas Health Center, Tyler, Te-xas 757101;Division of Bacterial Diseases, Tuberculosis and other Mycobacteriosis Laboratory Section,

National Center for Infectious Diseases, Centers for Disease Control and Prevention,Atlanta, Georgia 303332; and the National Chubu Hospital, Obu-Aichi, Japan3

Received 21 April 1993/Returned for modification 3 June 1993/Accepted 1 September 1993

Mycobacterium chelonae-like organisms are nonpigmented rapidly growing mycobacteria whose clinicalsignificance is unknown. We evaluated 87 sporadic isolates encountered in a clinical laboratory. Most isolates(62%) were respiratory; only 2 of 54 (4%) (both from patients with AIDS) were clinically significant. Among33 nonrespiratory isolates, 20 of 33 (or 61%) were clinically significant. Clinical diseases included posttrau-matic wound infections and catheter-related sepsis. Routine biochemical features included growth inhibition by5% NaCl (100o), a smooth colony morphology (94%), positive 3-day arylsulfatase reaction (84%), no color ora light tan color on iron uptake (100%Yo), and variable nitrate reduction (45%). Additional characteristics thathelped to separate this group from M. chelonae and Mycobacterium abscessus were susceptibility to cephalothin(90%) and ciprofloxacin (100o), utilization of mannitol (94%) and citrate (83%) as carbon sources, and uniquepatterns of mycolic acid esters by high-performance liquid chromatography. This group was quite drugsusceptible, with 100lo of isolates inhibited by amikacin, imipenem, cefoxitin, cefmetazole, and the newerquinolones ciprofloxacin and ofloxacin. Three examples of this group, including a proposed type strain, havebeen deposited in the American Type Culture Collection.

In 1982 an outbreak of peritonitis that involved the use ofhospital-based automated chronic peritoneal dialysis ma-chines was reported. Both epidemic and sporadic casesoccurred over a 2-year period, with two separate dialysiscenters being involved. The causative organism, a previ-ously unrecognized nonpigmented species of rapidly grow-ing mycobacteria, was recovered from tap water in thedialysis center as well as the dialysis machines (1). Theorganism was similar to Mycobacterium chelonae (formerlyM. chelonae subsp. chelonae) in that it was arylsulfatasepositive at 3 days, nonpigmented, and nitrate negative anddid not grow in the presence of 5% NaCl. The organism wasgiven the tentative name of M. chelonae-like organism(MCLO) (1, 18). Although the isolates were uniform in theirbiochemical reactions (1) and antimicrobial susceptibilitypatterns (20), subsequent evaluation including studies of13-lactamase electrophoretic patterns (29) suggested that asingle strain of organism may have been involved, andhence, the phenotypes of sporadic isolates may vary fromthis initial description.

Since the report of the peritonitis outbreak, we haveencountered sporadic isolates of MCLOs among isolates ofrapidly growing mycobacteria submitted for susceptibilitytesting. The purpose of the current study was to characterizethe phenotypic features of these strains, their clinical signif-icance, and their antimicrobial susceptibility patterns.

* Corresponding author.t Present address: Takaoka Clinic, Toyota Motor Corporation,

Honda-cho, Toyota, Japan 473.

MATERIALS AND METHODS

Isolates of rapidly growing mycobacteria with features ofthe MCLO were identified among clinical and environmentalisolates submitted to the Nocardia/Mycobacterial ResearchLaboratory of the University of Texas for susceptibilitytesting. Most isolates were then characterized biochemicallyand for growth features at the Mycobacterial ReferenceSection of the Centers for Disease Control and Prevention,Atlanta, Ga., by previously described methods (18). Se-lected isolates were also submitted to the National ChubuHospital, Obu-Aichi, Japan, for testing against 103 charac-ters that included growth characteristics, utilization of vari-ous carbohydrates as carbon and/or nitrogen sources, andgrowth inhibition by a variety of drugs and chemical com-pounds (22).

Isolates were tested for their antimicrobial susceptibilitiesby the broth microdilution method (at the Texas laboratory)in cation-supplemented Mueller-Hinton broth as describedpreviously for rapidly growing mycobacteria (20). Isolateswere incubated at 30°C for 72 h in room air. MICs wereinterpreted as the lowest concentration with no visiblegrowth, except for sulfamethoxazole, for which 80% orgreater inhibition of growth was used. Isolates were alsotested for susceptibility by disk diffusion (Texas laboratory)by using Mueller-Hinton agar and commercial drug disks ofpolymyxin B (300 U) and cephalothin (30 ,g) (BBL, Cock-eysville, Md.) and pipemidic acid disks (5 jig) (obtained fromthe Mycobacteriology Section, Centers for Disease Controland Prevention) with zones of partial or complete inhibition(pipemidic acid and polymyxin B) or complete inhibitiononly (cephalothin) recorded after 72 h of incubation at 30°Cin room air (7, 24). Control strains for broth microdilutionwere Escherichia coli ATCC 25922 and Mycobacterium

3231

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

3232 WALLACE ET AL.

fortuitum ATCC 6841, while the control strains for agar diskdiffusion were E. coli ATCC 25922 and Staphylococcusaureus ATCC 25923.As a control for the cephalothin disk test, 20 isolates each

of M. chelonae (formerly M. chelonae subsp. chelonae) (13),Mycobacterium abscessus (formerly M. chelonae subsp.abscessus) (13), M. fortuitum, Mycobacterium peregninum(13), M. fortuitum third biovariant complex (10 isolates eachbeing sorbitol positive and sorbitol negative), and Mycobac-terium smegmatis identified by standard published methods(7, 24, 25) were also tested. The type strain for each of thesix groups (obtained from the American Type Culture Col-lection, Rockville, Md.) was included in the test group.High-performance liquid chromatography (HPLC) of my-

colic acid esters was done by two techniques, both of whichare well characterized elsewhere (8). The initial techniqueused various flow rates and solvent concentrations of chlo-roform and methanol, with a run time of 42 min. The mostrecently adopted technique, which is a modified version of amethod developed at the Texas Department of Health,Austin, used methylene chloride and methanol as the sol-vents, with a constant flow rate and a run time of only 8 min.Organisms were grown and processed under fixed conditionsas described previously (8). A high-molecular-weight inter-nal standard for HPLC (Ribi ImmunoChem Research, Inc.,Hamilton, Mont.) was used as a reference standard.

RESULTS

Organisms were included as isolates of MCLO if theygrew in less than 7 days at 30°C, were nonpigmented, wouldnot grow in 5% NaCl at 30°C after 2 weeks, had a negativeiron uptake but with (usually) a light tan or rust colorationalong the edge of the slant, and were highly drug susceptiblein a pattern not seen with M. chelonae or M. abscessus (20).A total of 87 isolates meeting this criteria were identified.The first 64 of these isolates were chosen for evaluation atthe Tuberculosis and other Mycobacterosis Laboratory Sec-tion of the Centers for Disease Control and Prevention. Allisolates met the MCLO inclusion criteria. (Because of theaccuracy of the screening criteria, the last 23 isolates werenot submitted for confirmation.) The isolates were uniformin a number of other biochemical or growth responses (Table1), with approximately 90% of strains having a low semi-quantitative catalase reaction (<45 mm), a negative 68°Ccatalase reaction, a positive 3-day arylsulfatase reaction,growth on MacConkey agar without crystal violet (manyisolates produced a filmy growth on MacConkey agar ratherthan the mature, heavy growth often seen with isolates of theM. fortuitum complex), utilization of mannitol and citratebut not inositol as carbon sources, and a smooth, almostmucoid, colonial morphology. A small number of strainswere negative for one of these features, but otherwise theylooked typical. This included some strains which would notgrow on MacConkey agar and others with a negative 3-dayarylsulfatase reaction. The feature with the most variabilitywas nitrate reduction (45% of strains).

Fifty-three of the isolates submitted to the Centers forDisease Control and Prevention were also tested by usingthe 103 characters used by Tsukamura (22). These isolatesgave homogeneous responses, but with no additional test

being identified that would separate the MCLO from M.fortuitum, M. chelonae, or M. abscessus. Selected results ofthese tests are shown in Table 1. A comparison of MCLOwith other nonpigmented rapidly growing mycobacteria isshown in Table 2.

Antimicrobial susceptibility testing was performed on all87 isolates for 10 antimicrobial agents by broth microdilutionMICs. Because MICs were determined at the time thatisolates were sent to the laboratory and because some drugswithin the MIC panels changed, MICs of seven additionaldrugs were determined against a smaller number of isolates(ranges, 10 to 59 isolates). The MICs of 11 of the drugs(amikacin, kanamycin, gentamicin, tobramycin, cefoxitin,cefmetazole, imipenem, ampicillin, amoxicillin-clavulanicacid, ciprofloxacin, and ofloxacin) generally fell within avery narrow range, with the mode and MICs for 50% and90% of isolates tested being within 1 dilution of each other(Table 3). Erythromycin and sulfamethoxazole had a widerange of MICs, but results for both drugs were generallydifficult to read because of a trailing endpoint. The only drugthat had easily readable endpoints in the broth system and adefinite bimodal MIC pattern was doxycycline; the MIC ofdoxycycline for 26 isolates (31%) was <0.25 jxg/ml and thatfor 37 isolates (44%) was 216 ,ug/ml (resistance). The MICsof clarithromycin (6 isolates) (3), ciprofloxacin (62 isolates),and ofloxacin (16 isolates) have been reported previously(24).MCLOs were unusual for mycobacteria in their suscepti-

bility to ampicillin and cephalothin, a finding noted in thefirst study of MCLO antimicrobial susceptibilities reportedby Swenson et al. (20) but not seen with other nonpigmentedrapidly growing mycobacteria (20). A disk diffusion suscep-tibility test with cephalothin was added to the diagnostictests used for MCLOs. By using a 20-mm or greater zone ofcomplete growth inhibition with the 30-Rg commercial drugdisk as the breakpoint, 90% of the 84 strains of MCLO werefound to be susceptible to cephalothin. Eight strains (10%),including the proposed type strain ATCC 49649, had zonesof 10 mm or less, four strains (5%) had zones of 20 to 29 mm,while the remainder (85%) had zones of 30 mm or greater.Eighty clinical isolates submitted to the Texas laboratory forsusceptibility testing, representing all of the other groups ofnonpigmented rapidly growing mycobacteria encountered inthe clinical laboratory, were also tested against cephalothin.Although some strains of M. smegmatis produce a latepigment, many do not, and hence, this species is one of thegroups included for comparison (26). None of the 80 clinicalstrains or the 6 strains from the American Type CultureCollection (representing six taxonomic groups) had any zoneof inhibition, defining the high specificity of cephalothinsusceptibility for identifying isolates of MCLOs.Three examples of MCLOs were deposited in the Ameri-

can Type Culture Collection. ATCC 49649 was an isolatefrom a patient with peritonitis and from the first describeddisease outbreak in Washington State in 1977 (1). It is theearliest characterized isolate of this taxonomic group and isproposed as the type strain. It is somewhat atypical in that itis cephalothin resistant and has high (>45 mm) semiquanti-tative catalase activity. Two other, more typical strains(ATCC 49650 and 49651) are also included. ATCC 49650 wasfrom a neck abscess and was chosen as an example of anitrate-positive strain. ATCC 49651 was from a postinjectionbuttock abscess. Both of the last two strains were fromTexas. The biochemical features of these strains are in-cluded in Table 1, and the MICs for the strains are given inTable 3.

All isolates submitted to the Centers for Disease Controland Prevention laboratory (64 strains) were studied byHPLC. The HPLC patterns of mycolic acid esters of the newspecies provided an easy means of separating them fromother nonpigmented, rapidly growing mycobacterial species.

J. CLIN. MICROBIOL.

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

M. CHELONAE-LIKE ORGANISM 3233

TABLE 1. Biochemical and growth characteristics of MCLOs

No. of Reaction of Reaction of the following ATCC strain:Characteristic isolates tested isolates

% Reacting49649 49650 49651

Growth within 7 days 87 + 100 + + +

Pigmentation

Smooth colony type

Growth at:280C300C370C420C450C520C

Arylsulfatase reaction3 days14 days

Iron uptake

Nitrate reduction

Catalase productionSemiquantitative (>45mm)At 68°C, pH 7

Substrates as carbon sourcesCitrateInositolMannitolSorbitolTrehalose

Growth on MacConkey agarb

Tolerance to 5% NaCl

Growth inhibitionPicrate (0.2%)NH20H (500 p.g/ml)Isoniazid (10 ,ug/ml)Rifampin (25 ,ug/ml)Ethambutol (5 pg/ml)Cephalothin (30-,ug disk)Polymyxin B (300-U disk)Pipemidic acid (5-,ug disk)

Acetamidase

Tween hydrolysis

87

64

498787494949

64

87

64

64

6464645353

64

87

4953535353878787

53

54

0

+

+++

-a+4

94

100100100

640

8491

0

+

+++

+

+ +

+++ +

+ +

+

45

1512

+

830

94210

89

+ +

+

+ +

0

+

1001300

32906111

+ ++

91

93

+

+ + +

a Most isolates produced a light tan color or a darker rust color along the edge of the slant but not the dark brown appearance seen with M. fortuitum.b Without crystal violet.

The majority of isolates had two sets of peaks, one of whichextracted earlier than any of the major peaks seen with M.fortuitum, M. chelonae, or M. abscessus (Fig. 1). By therapid (8-min) extraction method, these peaks were seen atbetween 2.3 and 3.2 min. These peaks, as well as a secondset of peaks that came off between 4.6 and 5.5 min (by therapid method), were present in all strains. The second peaksoccurred at the same time as the peaks seen in most otherrapidly growing mycobacterial species. A comparison oftypical HPLC patterns of this type as exhibited by ATCC49649 (type strain) and ATCC 49650 and typical patterns forM. fortuitum and M. abscessus are shown in Fig. 1. A

second pattern of mycolic acid esters was exhibited by asmall number of isolates of MCLOs. These strains had athird set of major peaks that occurred between the first two.An evaluation of the biochemical features and drug suscep-tibilities revealed no unique features of the isolates with thissecond HPLC pattern.

Clinical information was available for all 87 isolates (Table4). The most common source was respiratory (62%). Onlytwo of these samples (4%) were thought to be clinicallysignificant (on the basis of the recovery of one of the isolatesfrom an open lung biopsy and the presence of the other inboth blood and sputum). Interestingly, both of these patients

VOL. 31, 1993

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

3234 WALLACE ET AL.

TABLE 2. Common characteristics of clinically significant nonpigmented rapidly growing mycobacteriaaM. fortuitum

Test MCLO M. chelonae M. abscessus M. fortuitum M. peregrinum M. smegmatis biovar IIIGroup 1 Group 2

Growth in 7 days +b + + + + + + +

Pigmentation - - - - - +

Growth at:300C + + + + + + + +370C + + + + + + + +450C - - - + + -

Arylsulfatase reaction3 days + + + + + + +14 days + + + + + + + +

Iron uptake Tan - - + + + + +

Nitrate reduction - - - + + + + +

Growth on MacConkey agar + + + + + + + +

Substrates as carbon sourcesMannitol + - - - + + + +Inositol - - - - - + + +Citrate + + _ _ - -Sorbitol - - _ _ - + +

Tolerance to 5% NaCl - - + + + + + +

Disk susceptibilityPolymyxin B + - - + + + + +Cephalothin +

HPLC pattern M C C, A F F S, F F Fa Summarized from references 1, 11-13, 18, 25-27.b Unless otherwise stated, all testing was done at 300C. +, 80% or more positive reactions; -, 80% or more negative reactions; +, intermediate values.c M, MCLO profile; C, M. chelonae profile; A, M. abscessus profile; F, M. fortuitum profile; S, M. smegmatis profile.

had AIDS. The remaining 52 respiratory isolates were singlepositive cultures from patients without evidence of myco-bacterial diseases or no available history. None were knownto have had a positive acid-fast bacillus smear. The nextmost common source was skin or soft tissue (14 isolates;16%). The most significant, isolates were seen followingaccidental trauma or surgery. The third most commonsource (15% of strains) was blood cultures. Eight patientswith long-term venous catheters and either underlying ma-lignancy (seven patients) or AIDS (one patient) developedcatheter sepsis. All had multiple positive blood cultures, andtwo had positive catheter tip cultures. One patient withAIDS had pneumonia with positive sputum and blood cul-tures (also included as a significant respiratory case). Theremaining five isolates were from patients with single posi-tive blood cultures with other blood cultures that werenegative. None had long-term venous catheters, none had anapparent mycobacterial disease, and none were treated.These five isolates were considered contaminants. Otherclinical sources were urine (two isolates), peritoneal fluid(two isolates), and cerebrospinal fluid (two isolates). Over-all, 20 of the 87 isolates (23%) were known to representdefinite clinical disease, and another two isolates (one peri-toneal, one cutaneous) (2%) were likely to be significant onthe basis of their sources, but patient histories were inade-quate. Overall, the incidence of significance among respira-

tory isolates was 2 of 54 (or 4%), in comparison with 20 of 33(65%) among nonrespiratory isolates. A review of the 20patients with definite disease is shown in Table 5.

DISCUSSION

The eponym MCLO was given to the organisms describedhere because of their similarity to M. chelonae (formerly M.chelonae subsp. chelonae) (13) in that they are nitratenegative, are inhibited by 5% NaCl, and fail to turn darkbrown when grown on ferric ammonium citrate (iron uptake)(1). The name is now misleading, because almost 50% ofthese sporadic strains reduce nitrate (like M. fortuitum, M.peregnnum, and M. smegmatis), and the strains are nottotally negative on iron uptake because they often produce alight tan or rust color along the edge of the slant, whereas M.chelonae does not.A small percentage of strains (14%) of this group had

negative 3-day arylsulfatase reactions, and an even smallerpercentage (6%) had a negative 14-day reaction. Thesestrains were, by other tests, similar to other members of thisgroup. Until more weighted taxonomic tests (DNA-DNAhybridization or 16S rRNA sequencing) are performed onthese isolates to determine whether these differences relateto significant genetic differences, we have elected to keepthem as part of the MCLO species.

J. CLIN. MICROBIOL.

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

M. CHELONAE-LIKE ORGANISM 3235

TABLE 3. Susceptibilities of MCLOs to 15 antimicrobial agents

MIC (>Lgftl)'Agent N.oisltsATCC straintested 50% 90% Range Mode

49649 49650 49651

Amikacin 84 1 4 .0.25-4 1, 2 4, 8 <0.5, <0.25 1, 1Kanamycin 84 2 4 s0.25-8 2 1 2Gentamicin 30 8 16 2-8 8 8 4 8Tobramycin 84 8 16 2-32 8 8, 16 4 8Streptomycin 10 4 8 4-16 4 8 4 4Cephalothin 27 4 32 1->64 2 128 16 4Cefoxitin 84 8 16 2-16 8 8, 16 8 8Cefmetazole 84 8 8 2-16 8 8 4 8Imipenem 84 1 2 .0.25-4 1 2, 2 1, <0.5 4, <0.5Ampicillin 33 2 4 .0.25-16 2 4, 2 1 2Amox-clav acidb 59 4 16 <0.5-32 4 1 0.5 4,2Erythromycin 84 1 8 .0.25->16 0.5, 1 1, 0.5 1, <0.125 2, <0.125Clarithromycin 15 0.25 0.5 .0.063-2 0.25 <0.063 <0.125Sulfamethoxazole 84 16 >128 s1-> 128 16 16, 32 >128, 4 8, 16Ciprofloxacin 84 0.5 2 0.125-2 0.5 2 0.5 0.25Ofloxacin 16 2 4 0.25-4 4 2 1 1Doxycycline 84 8 >32 .0.25->32 .0.25 .0.25, .0.25 8 32

a 50% and 90%, MICs for 50 and 90% of isolates tested, respectively.b Amox-clav acid, amoxicillin-clavulanic acid ratio of 2:1.

The clinical data suggest that the organisms are mostcommonly recovered from sputum samples but are almostnever associated with lung disease. We recommend fol-low-up or performing drug susceptibility testing only if anisolate of this group is recovered from multiple sputumsamples or is from a patient with AIDS. (In the currentstudy, almost all respiratory isolates were from single posi-tive samples, with other samples being negative.)

In contrast, recovery of organisms of this taxonomicgroup from skin or blood cultures was often associated withclinical disease. The two syndromes of disease identifiedwere catheter-associated sepsis and posttraumatic woundinfections. Single positive blood cultures were often contam-inated, so careful attention to the clinical setting or thepresence of multiple positive blood cultures was required toconfirm pathogenicity. The types of disease seen with theseorganisms are typical for those caused by other rapidlygrowing mycobacterial species, including M. fortuitum, M.chelonae, M. abscessus, and M. smegmatis (25, 26, 28).

It was not possible to determine the incidence of thistaxonomic group among consecutive routine isolates submit-ted for susceptibility testing because, for part of the studyperiod, one of the major contributors of organisms to theTexas laboratory (Texas Department of Health, Austin) wasasked to submit all isolates that were suspected of belongingto this group. Most of the isolates submitted in this mannerwere from single positive sputum cultures (not considered tobe clinically significant) and, hence, would not normallyhave been sent for susceptibility testing. A review of allpreviously unidentifiable or atypical nonpigmented rapidlygrowing mycobacteria submitted to the Texas laboratorybetween 1982 and 1988 revealed that more than 95% of theseisolates were either isolates of the recently "rediscovered"species M. smegmatis (26) or this group, with the majoritybelonging to the MCLO group (23a). A review of all isolatesof nonpigmented rapidly growing mycobacteria submitted tothe Mycobacterial Reference Section of the Centers forDisease Control and Prevention for identification in 1991revealed that 20% were isolates of MCLOs. Reasons fororganism submission varied, but they usually resulted from

an atypical biochemical result (e.g., citrate negative) or aprobable lack of familiarity with the taxonomic group (7a).The latter two observations suggest that these organisms arenot an infrequent cause of problems with species identifica-tion, but they do not provide the true incidence of this groupamong consecutive clinical isolates.A major issue to be addressed is whether current standard

laboratory techniques will identify MCLOs. One key bio-chemical feature of this group is their growth inhibition in 5%NaCl. Usually, the inability to grow in the presence of 5%NaCl among isolates of rapidly growing mycobacteria hasbeen limited to isolates of M. chelonae. Unfortunately,isolates of M. abscessus are often inhibited unless lowtemperatures are used (28°C) and incubation is done for up to4 weeks (11). Thus, growth inhibition by 5% NaCl is not ahighly specific indicator of an MCLO. It will, however,exclude nonpigmented rapidly growing species other thanM. chelonae or M. abscessus, because other species growreadily on 5% NaCl (18, 25, 26).When recognized, the light tan or rust color of the colonies

on ferric ammonium citrate is highly suggestive of MCLObecause this characteristic has not been seen with otherrapidly growing groups. This pattern is not always present orrecognized, however. The remaining tests currently per-formed to define the species (18) are inadequate for recog-nition of the MCLO. Thus, additional laboratory tests areneeded.Three additional techniques which will provide for identi-

fication are antimicrobial susceptibility tests, carbohydrateutilization tests, and the HPLC patterns of mycolic acidesters. Susceptibility studies have not been routinely usedfor species or subgroup identification with mycobacteria.They have frequently been used in laboratories that performboth susceptibility testing and mycobacterial identification.Susceptibilities to polymyxin B (27), amikacin (27), cipro-floxacin (19, 24), ofloxacin (23), and pipemidic acid (6, 19)have all been suggested for separating M. fortuitum from M.chelonae and M. abscessus. Susceptibility test results arehelpful in separating M. chelonae and M. abscessus from theMCLOs, because M. chelonae and M. abscessus are quite

VOL. 31, 1993

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

3236 WALLACE ET AL.

M. abscessus

(min.)

MCLO (ATCC 49649)(type strain) C

STD

A

00n

co

n

.0

a)

Q

C)C.CucoI-0(0.0Cu

a

C!

8a9

M. fortuitum B

89

a

(min.)

MCLO (ATCC 49650) D

STD

.

9i

99t. 99~u99.9. 9999

C! CS CS

I

Cq85 99 us 99#C.) la tn t 04 e C4 " to leX

w

(min.) (min.)FIG. 1. HPLC chromatograms of a clinical strain of M. abscessus (A), a clinical strain of M. fortuitum (B), and two ATCC strains of

MCLOs (C and D). Std, high-molecular-weight standard.

drug resistant, while the MCLOs are relatively drug sus-

ceptible. Susceptibility to cephalothin, ampicillin, amoxi-cillin-clavulanic acid, doxycycline, ciprofloxacin, and sulfa-methoxazole is common among MCLOs, but more than 95%of M. abscessus isolates are resistant to these drugs (20, 21,24). Susceptibilities to two agents (ciprofloxacin and ceph-alothin) by the disk diffusion method appeared to be mostuseful taxonomically. Susceptibility to ciprofloxacin wouldreadily separate MCLOs from M. chelonae or M. abscessusbut not from M. fortuitum. Susceptibility to cephalothin wasthe most specific test, because M. fortuitum, M. peregrinum,M. abscessus, and M. chelonae are resistant to cephalothin,approximately 90% of strains of MCLOs are susceptible tothe drug.The second group of tests that could be used to identify

the MCLO is the use of various carbohydrates as carbonsources (with or without acid production). Carbohydrateutilization tests for the identification of rapidly growingmycobacteria have been performed previously, but only by asmall number of reference laboratories. Its usefulness has

included separation ofM. chelonae subspecies (M. chelonaesubsp. chelonae is citrate positive; M. chelonae subsp.abscessus is citrate negative) (18). Recent recommendationsthat M. chelonae and M. abscessus be designated as sepa-rate species (13, 14) will require testing of their abilities toutilize citrate, given the problems with the use of growthinhibition by 5% NaCl, as noted above. Recent studies basedon DNA-DNA hybridization (13, 14) have also suggestedthat M. fortuitum biovariant peregninum be returned tospecies status, hence, its designation in this report as M.peregrinum and the designation of M. fortuitum biovarfortuitum as M. fortuitum. Carbohydrate utilization is essen-tial for separation of M. peregrinum from M. fortuitum (M.peregrinum is mannitol positive, while M. fortuitum ismannitol negative) (18) because other routinely used bio-chemical tests and HPLC will not separate the two species.Recent recognition of another taxonomic group of yet un-

proven but likely species status is M. fortuitum third bio-variant complex (25). These isolates behave biochemicallylike M. fortuitum, except they are positive for mannitol and

a)U

coCL.00(0.0Cu

3.99 4.27 4.56 4.85 5.13 5.42

a)C.)c

L.00(0)co.0

8

J. CLIN. MICROBIOL.

cm Cl) N"It 9 . . . 4 1: 1:In Cl) V) -0 * qp 0 In

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

M. CHELONAE-LIKE ORGANISM 3237

TABLE 4. Clinical sources of 87 isolates of MCLOs obtainedfrom 84 patients

Source Total no. No. of isolates knownSource of isolates to be significant

Respiratory 54 2Sputum 37 la

Bronchial wash 8 0Pleural fluid 4 0Gastric aspirate 2 0Open lung biopsy 1 1Transbronchial lung biopsy 1 0

Skin or soft tissue 16 10Leg 4 2Lymph node 4 2Miscellaneousb,c 5 3Arm 3 3

Blooda, b 11 7

Urine 2 0

Cerebrospinal fluid 2 1

Peritoneal fluid 2 1

Total 87 (84 patients) 21 (18 patients)a One patient had positive sputa and positive blood cultures and is included

in both groups.b Two patients had positive catheter site cultures and positive blood

cultures and are included in both groups.c Miscellaneous includes catheter sites (n = 2), breast (n = 1), esophagus (n

= 1), and buttock (n = 1).

inositol (18, 25). For MCLOs, routine isolates are positivefor mannitol and citrate and negative for inositol, a patternnot identified among any other rapidly growing species orsubgroup.The third test which provided ready separation of the new

species from other rapidly growing nonpigmented mycobac-terial groups was HPLC of mycolic acid esters. Isolates of

MCLOs have a unique set of peaks that is not evident withthese other taxonomic groups that are evolving.We believe that the most useful single test overall is

carbohydrate utilization (mannitol, inositol, and citrate),because it identifies M. fortuitum, M. peregninum, M. ab-scessus, M. fortuitum third biovar complex, and M. chelo-nae and helps to identify isolates of M. smegmatis (26) aswell as allows for the identification of isolates of MCLOs. Ifthis test is not diagnostic (e.g., about 20% of isolates of thenew group will be mannitol positive but citrate negative),cephalothin susceptibility should be performed. Susceptibil-ity to cephalothin alone is also acceptable as the adoptedroutine test for recognition of the MCLOs, but its usefulness(unlike carbohydrate utilization) is limited to MCLOs. Ide-ally, both sets of tests should be performed. As HPLCbecomes more widely available, it may replace some or all ofthese prior tests.The ecologic niche of the MCLO taxonomic group in the

environment is unknown, because previous environmentalstudies had inadequate means of identification to recognizethem. Without question, these organisms are common inpotable water, more specifically, tap water. A number ofoutbreaks of nosocomial disease caused by the rapidlygrowing mycobacteria have been reported over the lastdecade. In reported outbreaks (2, 12, 16, 17) and severalunreported ones (23a) in which tap water was evaluated formycobacteria, this new group (MCLOs) has been recovered.This suggests that MCLOs may be universal in tap water,including the tap water in hospital water systems. Theorganism has been recognized as a nosocomial pathogenonly in the peritonitis outbreak in Washington State (1) andin the hemodialysis outbreaks in Louisiana (2) and California(16). In each situation, the use of tap water contaminatedwith the new species resulted in contamination of reusable orprocessed dialysis equipment.The frequent presence of this organism in tap water and

processed water probably reflects its presence in naturalwater supplies and its relative resistance to standard disin-fectants such as chlorine, formaldehyde, and glutaraldehyde(4, 5, 10). Fortunately, nosocomial disease as a consequenceof the presence of this species in hospital water systems is

TABLE 5. Clinical disease caused by MCLOsa

Patient no. Illness Positive cultures Underlying disease or condition

1 Catheter sepsis Multiple blood, catheter Malignancy2 Catheter sepsis Multiple blood Malignancy3 Catheter sepsis Multiple blood, catheter Malignancy4 Catheter sepsis Multiple blood Malignancy5 Catheter sepsis Multiple blood Malignancy6 Catheter sepsis Multiple blood AIDS7 Catheter sepsis Multiple blood Malignancy8 Catheter sepsis Multiple blood Malignancy9 Meningitis Multiple CSF samples (AFB smear positive) AIDS; prior LP for unrelated reason 3 wk previously10 Pneumonia Sputum, blood AIDS11 Pneumonia Open lung biopsy AIDS12 Subcutaneous abscess Abscess fluid (buttock) Prior intramuscular injections13 Cellulitis, osteomyelitis Wound, bone Open fracture, leg14 Cellulitis Surgical biopsy (hand) None15 Cellulitis Surgical biopsy (arm) (AFB smear positive) SLE on corticosteroids16 Cellulitis Surgical biopsy (wrist) Prior surgery17 Surgical wound infection Wound cultures (leg) Lower leg amputation18 Lymphadenitis Inguinal lymph node Adjacent trauma19 Lymphadenitis Cervical lymph node Infected thyroglossal duct cyst20 Peritonitis Peritoneal fluid Chronic renal failure on automated peritoneal dialysis

a Abbreviations: CSF, cerebrospinal fluid; LP, lumbar puncture; AFB, acid-fast bacillus; SLE, systemic lupus erythematosus.

VOL. 31, 1993

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

3238 WALLACE ET AL.

rare on the basis of previous reports of disease and thecurrent study. Its presence, however, may result in thecontamination or transient colonization seen in sputumsamples and is a potential source of contamination whenevertap water is used for laboratory studies or patient care.Proposed minimal standards for inclusion in the genus

Mycobacterium and for description of new Mycobacteriumspecies of slowly growing (but not rapidly growing) specieshas recently been published by L&vy-Frebault and Portaels(15). Previous studies with MCLOs demonstrated some of itsgrowth characteristics (1, 18). Guerrant et al. studied thehigh-temperature cleavage products of mycolic acids fromMCLOs by gas chromatography (9). As with other myco-bacteria, pyrolysis esters C22 to C26 were identified. Inter-estingly, this new group contains a unique mixture of C22 andC24 mycolic acid cleavage products, while other nonpig-mented rapidly growing species contain only C24 products.(The five strains of MCLOs studied were also examined inthe current study.) Most of the biochemical studies andmycolic acid profiles suggested for the slowly growingspecies have been added in the current study. Although theavailable phenotypic and biochemical data suggest thatMCLOs represent a distinct species, genetic relatednessstudies to other nonpigmented rapidly growing mycobacteriaare needed. Sequencing of selected regions of the 16S rRNAof the three MCLO strains from the American Type CultureCollection and DNA-DNA homology with other nonpig-mented rapidly growing mycobacteria are ongoing.

ACKNOWLEDGMENTNo support for this study was received from pharmaceutical

companies.

REFERENCES

1. Band, J. D., J. I. Ward, D. W. Fraser, N. J. Peterson, V. A.Silcox, R. C. Good, P. R. Ostrey, and J. Kennedy. 1982.Peritonitis due to a Mycobacterium chelonei-like organismassociated with intermittent chronic peritoneal dialysis. J. In-fect. Dis. 145:9-17.

2. Bolan, G., A. L. Reingold, L. A. Carson, V. A. Silcox, C. L.Woodley, P. S. Hayes, A. W. Hightower, L. McFarland, J. W.Brown III, N. J. Petersen, M. S. Favero, R. C. Good, and C. V.Broome. 1985. Infections with Mycobacterium chelonei in pa-tients receiving dialysis and using processed hemodialyzers. J.Infect. Dis. 152:1013-1019.

3. Brown, B. A., R. J. Wallace, Jr., G. 0. Onyi, V. DeRosas, andR. J. Wallace III. 1992. Activities of four macrolides, includingclarithromycin, against Mycobacterium fortuitum, Mycobacte-rium chelonae, and M. chelonae-like organisms. Antimicrob.Agents Chemother. 36:180-184.

4. Carson, L. A., L. A. Bland, L. B. Cusick, M. S. Favero, G. A.Bolan, A. L. Reingold, and R. C. Good. 1988. Prevalence ofnontuberculous mycobacteria in water supplies of hemodialysiscenters. Appl. Environ. Microbiol. 54:3122-3125.

5. Carson, L. A., L. B. Cusick, L. A. Bland, and M. S. Favero.1988. Efficacy of chemical dosing methods for isolating nontu-berculous mycobacteria from water supplies of dialysis centers.Appl. Environ. Microbiol. 54:1756-1760.

6. Casal, M. J., and F. C. Rodriguez. 1981. Simple, new test forrapid differentiation of the Mycobacterium fortuitum complex.J. Clin. Microbiol. 13:989-990.

7. Collins, C. H., M. D. Yates, and A. H. C. Uttley. 1985.Differentiation of Mycobacterium chelonei from M. fortuitumby ciprofloxacin susceptibility. J. Hyg. Camb. 95:619-621.

7a.Floyd, M., and V. Silcox. Unpublished data.8. Floyd, M. M., V. A. Silcox, W. D. Jones, Jr., W. R. Butler, and

J. 0. Kilburn. 1992. Separation of Mycobacterium tuberculosisand Mycobacterium bovis using high-performance liquid chro-

matography of mycolic acids. J. Clin. Microbiol. 30:1327-1330.9. Guerrant, G. O., M. A. Lambert, and C. W. Moss. 1981.

Gas-chromatographic analysis of mycolic acid cleavage prod-ucts in mycobacteria. J. Clin. Microbiol. 13:899-907.

10. Hayes, P. S., D. L. McGiboney, J. D. Band, and J. C. Feeley.1982. Resistance of Mycobacterium chelonei-like organisms toformaldehyde. Appl. Environ. Microbiol. 43:722-724.

11. Kubica, G. P. 1973. Differential identification of mycobacteria.VII. Key features for identification of clinically significantmycobacteria. Am. Rev. Respir. Dis. 107:9-21.

12. Kuritsky, J. N., M. G. Bullen, C. V. Broome, V. A. Silcox, R. C.Good, and R. J. Wallace, Jr. 1983. Sternal wound infections andendocarditis due to organisms of the Mycobacterium fortuitumcomplex. Ann. Intern. Med. 98:938-939.

13. Kusunoki, S., and T. Ezaki. 1992. Proposal of Mycobacteriumperegrinum sp. nov., nom. rev., and elevation of Mycobacte-num chelonae subsp. abscessus (Kubica et al.) to speciesstatus: Mycobactenum abscessus comb. nov. Int. J. Syst.Bacteriol. 42:240-245.

14. Levy-Frebault, V., F. Grimont, P. A. D. Grimont, and H. L.David. 1986. Deoxyribonucleic acid relatedness study of theMycobacterium fortuitum-Mycobacterium chelonae complex.Int. J. Syst. Bacteriol. 36:458-460.

15. Levy-Frebault, V., and F. Portaels. 1992. Proposed minimalstandards for the genus Mycobacterium and for description ofnew slowly growing Mycobacterium species. Int. J. Syst. Bac-teriol. 42:315-323.

16. Lowry, P. W., C. M. Beck-Sague, L. A. Bland, S. M. Aguero,M. J. Arduino, A. N. Minuth, R. A. Murray, J. M. Swenson, andW. R. Jarvis. 1990. Mycobactenum chelonae infection amongpatients receiving high-flux dialysis in a hemodialysis clinic inCalifornia. J. Infect. Dis. 161:85-90.

17. Lowry, P. W., W. R. Jarvis, A. D. Oberle, L. A. Bland, R.Silberman, J. A. Bocchini, Jr., H. D. Dean, J. M. Swenson, andR. J. Wallace, Jr. 1988. Mycobactenum chelonae causing otitismedia in an ear-nose-throat practice. N. Engl. J. Med. 319:978-982.

18. Silcox, V. A., R. C. Good, and M. M. Floyd. 1981. Identificationof clinically significant Mycobactenum fortuitum complex iso-lates. J. Clin. Microbiol. 14:686-691.

19. Steele, L. C., and R. J. Wallace, Jr. 1987. Ability of ciproflox-acin but not pipemidic acid to differentiate all three biovariantsof Mycobacterium fortuitum from Mycobacterium chelonae. J.Clin. Microbiol. 25:456-457.

20. Swenson, J. M., C. Thornsberry, and V. A. Silcox. 1982. Rapidlygrowing mycobacteria: testing of susceptibility to 34 anti-microbial agents by broth microdilution. Antimicrob. AgentsChemother. 22:186-192.

21. Swenson, J. M., R. J. Wallace, Jr., V. A. Silcox, and C.Thornsberry. 1985. Antimicrobial susceptibility of five sub-groups of Mycobacterium fortuitum and Mycobactenium chelo-nae. Antimicrob. Agents Chemother. 28:807-811.

22. Tsukamura, M. 1981. Numerical analysis of rapidly growing,nonphotochromogenic mycobacteria, including Mycobacteriumagni (Tsukamura 1972) Tsukamura sp. nov., nom. rev. Int. J.Syst. Bacteriol. 31:247-258.

23. Tsukamura, M. 1983. In vitro antimycobacterial activity of anew antibacterial substance DL-8280. Differentiation betweensome species of mycobacteria and related organisms by theDL-8280 susceptibility test. Microbiol. Immunol. 27:1129-1132.

23a.Wallace, R. J. Unpublished data.24. Wallace, R. J., Jr., G. Bledsole, G. Sumter, C. V. Sanders, L. C.

Steele, B. A. Brown, J. Smith, and D. R. Graham. 1990.Activities of ciprofloxacin and ofloxacin against rapidly growingmycobacteria with demonstration of acquired resistance follow-ing single-drug therapy. Antimicrob. Agents Chemother. 34:65-70.

25. Wallace, R. J., Jr., B. A. Brown, V. A. Silcox, M. Tsukamura,D. R. Nash, L. C. Steele, V. A. Steingrube, J. Smith, G. Sumter,Y. Zhang, and Z. Blacklock. 1991. Clinical disease, drug suscep-tibility, and biochemical patterns of the unnamed third biovari-ant complex of Mycobacterium fortuitum. J. Infect. Dis. 163:598-603.

J. CLIN. MICROBIOL.

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

M. CHELONAE-LIKE ORGANISM 3239

26. Wallace, R. J., Jr., D. R. Nash, M. Tsukamura, Z. M. Blacklock,and V. A. Silcox. 1988. Human disease due to Mycobacteriumsmegmatis. J. Infect. Dis. 158:52-59.

27. Wallace, R. J., Jr., J. M. Swenson, V. A. Silcox, and R. C. Good.1982. Disk diffusion testing with polymyxin and amikacin fordifferentiation of Mycobacterium fortuitum and Mycobacteniumchelonei. J. Clin. Microbiol. 16:1003-1006.

28. Wallace, R. J., Jr., J. M. Swenson, V. A. Silcox, R. C. Good,J. A. Tschen, and M. S. Stone. 1983. Spectrum of disease due torapidly growing mycobacteria. Rev. Infect. Dis. 5:657-679.

29. Zhang, Y., R. J. Wallace, Jr., V. A. Steingrube, B. A. Brown, R.Nash, A. Silcox, and M. Tsukamura. 1993. Isoelectric focusingpatterns of P-lactamases in the rapidly growing mycobacteria.Tubercle 73:337-344.

VOL. 31, 1993

on April 30, 2021 by guest

http://jcm.asm

.org/D

ownloaded from